Sampling Inspection with Classification Errors

Abstract

For complex multifunctional repairable products there are created functioning mathematical models of sampling inspection that enable to evaluate control efficiency in various situations. During such sampling, series of products, which were recognized as defective, are examined in complete (continuous) control, where defective products are repaired and again returned to the control. Within alternative control of complex products, beside right decisions about product quality there are taken wrong decisions, i.e.: good products are rejected as defective (sorted out), and defective products are recognized as good ones. In the result separate series and single products are classified with I and II sort errors. Probability values of these errors may grow significantly and result decrease of control efficiency and increase of losses. There are analyzed situations when sample products are classified without and errors, and within sampling there exists undeniable (evident) errors.